LAUSR

In recent times, medical scan images are crucial for accurate diagnosis by medical professionals. Due to the increasing size of the medical images, transfer and storage of images require huge bandwidth and storage space, and hence needs compression. In this paper, multilevel thresholding using 2-D histogram is proposed for compressing the images. In the proposed work, hybridization of optimization techniques viz., Genetic Algorithm (GA), Particle Swarm Optimization (PSO) and Symbiotic Organisms Search (SOS) is used to optimize the multilevel thresholding process by assuming the Renyi entropy as an objective function. Meaningful clusters are possible with optimal threshold values, which lead to better image compression. For performance evaluation, the proposed work has been examined on six Magnetic Resonance (MR) images of brain and compared with individual optimization techniques as well as with 1-D histogram. Recent study reveals that peak signal to noise ratio (PSNR) fail in measuring the visual quality of reconstructed image because of mismatch with the objective mean opinion scores (MOS). So, we incorporate weighted PSNR (WPSNR) and visual PSNR (VPSNR) as performance measuring parameters of the proposed method. Experimental results reveal that hGAPSO-SOS method can be accurately and efficiently used in problem of multilevel thresholding for image compression.